Lessons from Chile on tackling poverty

Productive inclusion is the buzzword taking shape in social policy circles in Latin America, and other middle income countries. Graduation out of social assistance does not equate with (or presume) a sustained exit from poverty.

As many middle-income countries are moving towards embracing cash transfers with or without co-responsibilities attached (and the recent hype of handing cash directly to the poor), there is an important wave of programs that provide “cash plus” intervention.

The cradle region where conditional cash transfers were born is embracing a new interesting shift towards integrated programs that expand the scope of traditional CCTs, especially for the extreme poor and vulnerable. There is a growing recognition that the provision of cash without consideration given to the importance of skills for income generation is not going to provide a sustained exit from poverty. Mexico and Brazil are shifting the design of their conditional cash transfer to place a broader emphasis on inclusion, with productivity and financial inclusion at the forefront.

Integrated and sequenced packages of programs targeted at the extreme poor have shown promising results in low income settings in pilot cases, where ‘ultra poor programs’ are being rolled out (see here). These ultra poor programs are carefully implemented, and the supply side of the integrated programs are generally vertically integrated within a given NGO. Can these programs deliver at scale? Can they be carried out within a system of interventions delivered by different agencies/ministries/service providers? For what objective and under what conditions?

Chile Solidario (CS) was the first of such large scale government programs explicitly targeted to the extreme poor. It combined demand side components for promoting access to transfers and services in the short term with supply side interventions to sustain exit from poverty through building skills and endowments. In the longer term the program was implemented at scale between 2002 and 2009.

Here are some design features of Chile Solidario that makes it stand out from traditional social protection programs:

• Identifying the needs of households and tailoring services to their needs: Information, transaction cost and psychological barriers are important obstacles to take-up of programs (see the seminal review by J. Currie here). These barriers are likely to compound for the poorest, who are exposed to multiple stressors in their daily lives. Social programs have evolved to target specifically these hard-to-reach target populations. The social or case workers work with the family to identify deficits along many welfare dimensions. The identification of needs can be done by a case worker through diagnostic visits (see reviews from Latin America here and here). In more complex and sophisticated settings where the human resources are available, as in Chile, social intermediation took the form of an intervention itself, with intensive psychosocial support through regular home visits of social workers to participating households for a period of two years.  The recognition of the psychosocial wellbeing as a key household asset has recently been eloquently exposed by Duflo in her Tanner lectures, and the psychology of poverty brought up in the behavioral economics literature has now taken the central stage in development economics (as  recently summarized in the WDR 2015). Qualitative feedback from program beneficiaries valued the aspect of being ‘made visible’, listened to, and engaged in a process that improves family cohesion and self-esteem. Interestingly, in depth qualitative work shows that the psychosocial benefits reported in the short run are perceived as instrumental to improvements in other dimensions in the medium term.
• The system approach – bridging the demand and the supply side of services:  The supply side component of the program aims at ensuring coordination among different programs, making the existing supply side available or creating new supply when needed. The rationale is to articulate the supply side to target the needs of a particular family. In the case of Chile, the program started in 2002, but the supply side reorganization and expansion became operational only after 2004, when the program became law and the institutional and budget arrangements between ministries and service providers could be put in place.

The nature of personalized interventions poses important challenges from an evaluation perspective as (i) the set of initial deprivations along different dimensions and the set of potential outcomes of interest is very different across participants and (ii) the set of programs made available to each participant is highly heterogeneous, and the articulation of the supply side took time to materialize.

The long term evaluation of the program (in this paper joint with Pedro Carneiro and Rita Ginja) relied on administrative data with a regression discontinuity design, exploiting the fact that the program eligibility was set to be a discontinuous function of the proxy means test. The attractiveness of the setting is that the threshold of eligibility varied across space by municipality, and over time by cohort, hence making the evidence less restrictive in terms of external validity than in the general RDD approach. The downside of using administrative data was that we could focus on only a few outcomes. What did we find?

Social protection: reaching the poorest and most socially excluded

• Demand activation: The program was extremely effective in achieving its social protection objective, activating the demand for a whole array of social services: the results hold for both the short and the long run.
• Social inclusion: the average results masked very important heterogeneity based on initial conditions: conditioning on baseline outcomes, we find that the results on take-up were mostly driven by those households in extreme poverty who were the hardest to reach, i.e. the intended target population.

From social protection to productive inclusion?

Qualitative work suggested that employment (especially as if relates to having a stable source of income in the household) and housing are among the most important aspirations of participating families and are perceived as well as to be key structural factors preventing households to escape extreme poverty in a sustained way. Yet we did not find impact on these two dimensions.

A key mechanism we explore to explain the lack of average effects on final outcomes are the differences in supply side availability. We measure this by exploiting the timing of the rollout of the supply side component of the program. Whereas public housing programs were highly rationed, employment programs rapidly expanded as the program became law, through a combination of self-employment, wage subsidies, soft-skills training and education equivalency.  We find significant employment effects for the spouses only for those cohorts exposed to an expanded supply side.

What next?

Drawing the lessons from the experience is hugely important for moving forward the discourse on social protection in middle income countries. More research needs to be built into program design, early on as the reality and implementation in many countries are moving much faster than impact evaluation.

• More clarity on social policy’s objectives is needed. Social protection, social inclusion, poverty alleviation, a (sequenced) combination of the three? Bridging the demand and the supply of social services, especially for the poorest sections of the population, is a key social protection objective in itself. For some key vulnerable groups, income support might be needed even in the long run. For sustained poverty alleviation (now ‘productive inclusion’), income stabilization needs to be combined with a much larger effort to identify key skills, employment bottlenecks, and to align a whole array of employment programs to the endowments/skills of the target population. Self-employment is a risky investment for low income households and might not be suited for all households in extreme poverty. Income generation for the extreme poor might range from technical and soft-skills training, to labor intermediation and placement. More is needed to identify what are the key packages of programs that match the needs of the target population in middle income settings, rural/urban environments.
• The key role of the activation of the supply side and the key aspect of sequencing: The lesson from the experience of Chile is that the supply side followed the demand. Adapting existing services- or creating new services – for the hardest to reach/serve have services and programs were created to meet the specific objectives of the minimum conditions expected to be met for the extreme poor.  Knowledge about their needs and about their constrained choice environment is first order.
• What is the key value added and role of social intermediation or soft-life skills support? What is the frequency and timing of contact? And for whom? What are they key profile of case workers and social workers? The content and structure of the curriculum? The answer will depend on initial institutional set up and the objectives that the social intermediation wants to achieve (diagnostics, activation of the demand for programs, matching demand and supply, psychosocial support).
• Advances are needed on the measurement agenda for psychosocial outcomes in survey work. My colleague Jed posted a few important blogs (see here and here). More is needed, especially in the context of program evaluation, thinking about identifying and measuring psychological domains that can be changed in the context of the intervention and modeling their importance as mediators for transformational change.

This post first appeared on The World Bank’s Let’s Talk Development Blog. Publication does not imply endorsement of views by the World Economic Forum.

To keep up with Agenda subscribe to our weekly newsletter.

Author: Emanuela Galasso is a senior economist in the Poverty and Inequality Unit of the Development Research Group.